In the paper, the costs of fighting are shown to be larger than the benefits. The title of the paper is "Fighting as a profit maximising strategy in the National Hockey League: more evidence". Based on this evidence, one would imagine a rational owner would not be encouraging fighting.
The positive blood-lust effect is what many fans would expect. NHL is a physical contest. However, as Rockerbie explains, fighting is not the primary reason fans come to see the game. Unlike the situation in boxing or mixed martial arts, fighting in hockey is a bonus feature. Fans are more concerned about winning. If the fighting damages the chances of winning then it seems fans would prefer less fighting.
The players and coaches, as decision makers, do not feature prominently in Rockerbie 's paper. There is a mention of intimidation but it is bound up with the negative impact of numerical disadvantage. One wonders about the extent to which individuals or teams seek to use fighting and/or intimidation up to the point before getting dismissed. In those sports that I'm more familiar, fighting and other forms of aggression tends to have other triggers, e.g. early in a game to "put down a marker" or later in a game when players lose discipline when the game slips from their grasp.
A previous post on this blog examined the cost of a red card in the English Premier League (here). It suggests that the cost of red cards might be overstated. The analysis in that post is fairly basic compared to that in Rockerbie's paper. However, it would be nice to know the extent to which violence, or the threat of violence, is used as a rational strategy by players and teams in various sports.